Esophageal and gastric carcinomas are among the top ten causes of cancer death worldwide. These diseases are usually detected at advanced stages, when available treatments are not very effective. Earlier detection of these cancers has been shown to make a significant impact on patient outcome. Our preliminary studies have discovered biomarkers that can diagnose synchronous cancer from normal or premalignant tissues. We will extend these studies to confirm that these same biomarkers can predict future (metachronous) esophagogastric cancer risk. This application will have as its ultimate goal the translation of these early detection biomarkers into clinical validation within 5 years. Aim 1. In pilot cohort Phase I discovery studies using cDNA microarrays, to develop biomarkers distinguishing between normal or metaplastic tissues of patients with vs. without cancer. Aim 1.a. In a cohort of 20 patients without Barrett's or cancer, 20 with Barrett's only, and 20 with Barrett's adenocarcinoma, using the shrunken nearest centroids predictor (SNCP) model, to identify gene expression panels and individual gene biomarkers to distinguish between patients with and without esophageal cancer. Aim 1.b. In 20 patients without and 20 with gastric cancer, using the SNCP model, to identify gene expression panels and individual gene biomarkers distinguishing between patients with and without gastric cancer. Aim 1.c. To determine the positive predictive value (PPV) of all potential biomarkers identified in Aims 1.a. and 1.b. Aim 2. In a pilot cohort Phase II validation study, to confirm expression panels and individual genes identified in Aim 1 with quantitative RT-PCR (Q-PCR) using a capillary microfluidics station-based method. Aim 2.a. To perform analytical validation of Q-PCR by checking its sensitivity in detecting a positive signal, its dynamic range, and its reproducibility in repeat analyses. Aim 2.b. To perform microfluidics measurement of gene expression levels from paraffin-derived esophageal tissues matched to the 60 frozen specimens studied in Aim 1.a. Aim 2.c. To perform microfluidics measurements of gene expression levels from paraffin-derived gastric tissues matched to the 60 frozen specimens studied in Aim 1.b. Aim 3. In larger cohorts, to perform Phase III cross-sectional retrospective longitudinal validation studies using the QPCR method analytically validated in Aim 2. Aim 3.a. To perform a Phase III study (n=400) of 200 esophageal cancer and 200 Barrett's patients, including 127 patients from whom tissues were obtained prior to a cancer diagnosis. Aim 3.b. To perform a Phase III study (n=200) of 100 gastric cancer and 100 noncancer patients. Aim 3.c. To perform statistical analyses of data from Aims 3.a. and 3.b. in order to determine their validity and significance.